Agile Development to Support State Assessment Analyses

Intern Update #1

Brian Harrold and Erik Whitfield
June 20, 2025

Project Overview

Brian Harrold and Erik Whitfield

Description

Create a suite of R packages to:

  1. Support policy-relevant analysis of state educational data
  2. Develop reproducible R functions for common analytic tasks
  3. Use an iterative development process to build, test, and refine tools
  4. Empower state analysts and researchers to produce high-quality, replicable reports

Goal

Develop robust functionality to:

  • Facilitate routine analysis
  • Streamline standard analytic workflows
  • Enable deeper, exploratory data investigations
  • Generate clear, reproducible reports and presentations
  • Provide comprehensive supporting documentation

Workflow

  • R and RStudio – Programming and package development
  • GitHub – Version control and website hosting
  • Quarto – Website creation and dynamic report generation

cohortED package

Brian Harrold

Focus of the Package

  • Cohort Analysis for Educational Research
    • Demographics
    • Student Mobility
    • Longitudinal Analysis (tracking cohorts over time)
    • Cross-Sectional Analysis (comparing cohorts)

Vocabulary

  • Classifications: Leave, Join, Stay
  • From Cohort Analysis: Churn, Persistence

Ideal Dataset

  • Longitudinal Student-Level Data
    • Example: sgpData_Long from SGPdata package
    • Requires: Student ID, Grade Level, Academic Year
    • Optional Demographics, Subgroups, Performance, School/District, etc.

Enabling the Ideal

  • Helper Functions
    • make_achievement_levels()
    • make_proficiency_levels()
    • make_mobility()
    • make_first_entered()
    • standardize_grade()

Example Output

Visualizing

  • Plotting Functions
    • plot_distribution_mobility()
    • plot_alluvial_mobility()
    • plot_cohort2_proficiency()

Example Output

Analyzing

  • Analysis Functions
    • More to come…

Example Report

Persistent Challenges

  • Designing flexible function arguments
  • Creating cohesive, user-friendly functionality
  • Handling data limitations
  • Developing dynamic reporting capabilities

Next Steps

  • Build in more functionality and analysis features
  • Expand dynamic reporting capabilities
  • Test with additional datasets (as available)
  • Obtain feedback

Desired Feedback

  • Functionality: Are there any features you think would be useful to add?
  • Reporting: Are there any challenges from states that this package could better support?
  • Datasets: Do you know of any datasets that can be used to test and validate the package?

Resources